Aberrant S-nitrosylation mediates calcium-triggeredventricular arrhythmia in the intact heartMichael J. Cutlera, Bradley N. Plummera, Xiaoping Wana, Qi-An Sunb, Douglas Hessb, Haiyan Liua, Isabelle Deschenesa,David S. Rosenbauma, Jonathan S. Stamlerb, and Kenneth R. Lauritaa,1

aHeart and Vascular Research Center, MetroHealth Campus of Case Western Reserve University, Cleveland, OH 44109; and bInstitute for TransformativeMolecular Medicine and Department of Medicine, Harrington Discovery Institute, Case Western Reserve University and University Hospitals Case MedicalCenter, Cleveland, OH 44109

Edited by Solomon H. Snyder, The Johns Hopkins University School of Medicine, Baltimore, MD, and approved September 14, 2012 (received for reviewJune 27, 2012)

Nitric oxide (NO) derived from the activity of neuronal nitric oxidesynthase (NOS1) is involved in S-nitrosylation of key sarcoplasmicreticulum (SR) Ca2+ handling proteins. Deficient S-nitrosylation of thecardiac ryanodine receptor (RyR2) has a variable effect on SR Ca2+

leak/sparks in isolated myocytes, likely dependent on the underly-ing physiological state. It remains unknown, however, whethersuch molecular aberrancies are causally related to arrhythmogenesisin the intact heart. Here we show in the intact heart, reduced NOS1activity increased Ca2+-mediated ventricular arrhythmias only in thesetting of elevated myocardial [Ca2+]i. These arrhythmias arosefrom increased spontaneous SR Ca2+ release, resulting from a com-bination of decreased RyR2 S-nitrosylation (RyR2-SNO) and in-creased RyR2 oxidation (RyR-SOx) (i.e., increased reactive oxygenspecies (ROS) from xanthine oxidoreductase activity) and could besuppressed with xanthine oxidoreductase (XOR) inhibition (i.e., allo-purinol) or nitric oxide donors (i.e., S-nitrosoglutathione, GSNO).Surprisingly, we found evidence of NOS1 down-regulation of RyR2phosphorylation at the Ca2+/calmodulin-dependent protein kinase(CaMKII) site (S2814), suggesting molecular cross-talk betweennitrosylation and phosphorylation of RyR2. Finally, we show thatnitroso–redox imbalance due to decreased NOS1 activity sensitizesRyR2 to a severe arrhythmic phenotype by oxidative stress. Ourfindings suggest that nitroso–redox imbalance is an importantmechanism of ventricular arrhythmias in the intact heart under dis-ease conditions (i.e., elevated [Ca2+]i and oxidative stress), and thattherapies restoring nitroso–redox balance in the heart could pre-vent sudden arrhythmic death.

Nitric oxide (NO) is an important regulator of cardiac functionvia both the activation of cyclic guanosine monophosphate-

dependent signaling pathways and direct posttranslational mod-ification of protein thiols (S-nitrosylation) (1). NO derived fromthe activity of neuronal nitric oxide synthase (NOS1) is involvedin S-nitrosylation of key sarcoplasmic reticulum (SR) Ca2+ han-dling proteins (2). In particular, nitrosylation of both skeletal andcardiac muscle ryanodine receptors (RyR1 and RyR2, respectively)alters their release properties, favoring activation (3, 4). Notably,an increase in RyR2 open probability can cause spontaneous SRCa2+ release, which may cause arrhythmias. Recently, it was shownthat decreased RyR2 S-nitrosylation (RyR2-SNO) through lossof NOS1, was associated with increased spontaneous SR Ca2+

release events in isolated cardiomyocytes, following rapid pacing(5). In a separate study, NOS1 deficiency was shown to decreasespontaneous SR Ca2+ sparks and the open probability of RyR2under resting conditions in cardiomyocytes and lipid bilayers,respectively (6). These studies suggest that NOS1 deficiencyhas a variable effect on RyR2 function, likely dependent on theunderlying physiological state (i.e., rapid heart rate versus qui-escence). It remains unknown, however, whether these changescreate a substrate for arrhythmogenesis in the intact heart.It is increasingly evident that activities of nitric oxide and re-

active oxygen species (ROS) are tightly coupled in cardiomyocytesproducing nitroso–redox balance. Elevated ROS production(oxidative stress) is a hallmark of several cardiovascular diseasesassociated with increased risk of fatal ventricular arrhythmias

[e.g., myocardial infarction (MI) and heart failure]. Burger et al.(7) recently demonstrated an increased incidence of ventriculararrhythmias following MI in NOS1-deficient mice. These datasuggest that a nitroso–redox imbalance may be arrhythmogenicin the setting of MI. However, the molecular basis of the increasedarrhythmogenesis is not known.In the current study, we found that decreased NOS1 activity

increased Ca2+-mediated ventricular arrhythmias only in thesetting of elevated myocardial [Ca2+]i. These arrhythmias arosefrom increased spontaneous SR Ca2+ release resulting from acombination of decreased RyR2-SNO and increased RyR2 oxi-dation (RyR2-SOx) [i.e., increased ROS from xanthine oxido-reductase (XOR) activity] and could be suppressed with xanthineoxidoreductase inhibition (i.e., allopurinol) or nitric oxide donors(i.e., GSNO). Notably, we found evidence of NOS1 regulation ofRyR2 phosphorylation at the Ca2+/calmodulin-dependent pro-tein kinase (CaMKII) site (S2814), suggesting molecular cross-talk between the nitrosylation and phosphorylation states ofRyR2. Finally, we show that nitroso–redox imbalance due todecreased NOS1 activity sensitizes RyR2 to a severe arrhythmicphenotype under oxidative stress.

ResultsEffect of NOS1 Inhibition in the Intact Heart on Susceptibility toVentricular Arrhythmia. In the intact heart, ventricular arrhyth-mias were seen following NOS1 inhibition only in the presence ofelevated [Ca2+]i. Fig. 1A shows an example of ECG and in-tracellular Ca2+ transient recordings from the intact heart underelevated [Ca2+]i conditions in the absence (Fig. 1A Upper) andpresence (Fig. 1A Lower) of NOS1 inhibition with S-methyl-L-thiocitrulline (SMTC). Following a halt in pacing (S1 − S1 = 190ms) under control conditions (without NOS1 inhibition) there isa spontaneous (nonelectrically driven) multicellular calcium re-lease (mSCR) event that was not sufficient to produce an ectopic(unstimulated) ventricular beat or arrhythmia (more than threeectopic beats). In contrast, in a heart treated with NOS1 inhibitor,multiple mSCRs (see arrows) developed that were associatedwith multiple ectopic beats (i.e., ventricular arrhythmia). Sum-mary data confirm that NOS1 inhibition increased susceptibilityto pacing-induced ventricular arrhythmias over a broad range ofpacing rates, P < 0.01 (Fig. 1B). However, this increased ar-rhythmia susceptibility following NOS1 inhibition was only evi-dent under elevated (n = 5) [Ca2+]i conditions. In fact, noventricular arrhythmias were seen with NOS1 inhibition undernormal [Ca2+]i conditions (n= 5) nor in control hearts under both

Author contributions: M.J.C., J.S.S., and K.R.L. designed research; M.J.C., B.N.P., X.W.,Q.-A.S., D.H., H.L., and I.D. performed research; J.S.S. contributed new reagents/analytictools; M.J.C., B.N.P., X.W., Q.-A.S., D.H., D.S.R., and K.R.L. analyzed data; and M.J.C., J.S.S.,and K.R.L. wrote the paper.

The authors declare no conflict of interest.

This article is a PNAS Direct Submission.1To whom correspondence should be addressed. E-mail: klaurita@metrohealth.org.

This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.1073/pnas.1210565109/-/DCSupplemental.

18186–18191 | PNAS | October 30, 2012 | vol. 109 | no. 44 www.pnas.org/cgi/doi/10.1073/pnas.1210565109

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0

normal (n = 4) and elevated (n = 4) [Ca2+]i conditions. Addi-tionally, single ectopic beats were seen under all conditions, yetthese were more likely to occur with NOS1 inhibition under el-evated [Ca2+]i conditions, P < 0.05. These data show that in well-

coupled myocardium, NOS1 inhibition and elevated myocardialCa2+ produce an arrhythmic substrate for triggered activity.

NOS1 Inhibition Combined with Ca2+ Overload Triggers Ca2+-MediatedArrhythmias. Previously, we have shown that the rate of mSCR riseis an important determinant of Ca2+-mediated triggered activity inintact heart preparations (8, 9). Therefore, to determine whetherthe arrhythmias we observed are due to mSCR activity followingNOS1 inhibition under elevated [Ca2+]i conditions, we comparedthe origin of consecutive ectopic beats with the location of maxi-mum mSCR slope. Fig. 2A, upper shows an example of a ventric-ular arrhythmia following NOS1 inhibition under elevated [Ca2+]iconditions. Shown are ECG and Ca2+ recordings from sites ofearliest (red trace) and latest (black trace) activation. After pacingwas terminated, the mSCR slope (dashed line) preceding the ec-topic ventricular beats V1, V2, and V4 is greater at the earliestactivated site compared with the latest activated site. Also, for thefirst beat (V1) the mSCR slope is greater and the preceding di-astolic interval (DI) is shorter compared with that for subsequentbeats. Contour maps of activation and mSCR slopes for V1–V4are shown (Fig. 2A, lower). In every beat except V2, the region ofearliest activation (red asterisk) corresponds exactly to the regionwhere the mSCR slope is greatest. V2 originated outside themapping field and may have been a reentrant beat. Importantly,because the endocardium was ablated, the site of earliest ac-tivity could not have occurred below the mapping field or fromPurkinje fibers. Thus, the earliest site of activation shown islikely the site where the ectopic beat originated.These findings were consistent across all preparations that de-

veloped ventricular arrhythmias. Additionally, NOS1 inhibition didnot prolong baseline QT interval compared with control (217 ±6 ms and 218 ± 6 ms, respectively; P = 0.9), suggesting that ectopywas not the result of early afterdepolarization. Thus, these datasuggest that ventricular arrhythmias measured in the intact heartduring NOS1 inhibition are Ca2+-mediated triggered arrhythmias.

Fig. 1. NOS1 inhibition (SMTC) during elevated [Ca2+]i increases suscepti-bility to triggered ventricular arrhythmias in the whole heart. (A) Example ofintracellular calcium recordings and ECG from the intact heart under ele-vated [Ca2+]i conditions in the absence (Upper) and presence (Lower) ofNOS1 inhibition. Following a halt in pacing (S1 − S1 = 190 ms) without NOS1inhibition (control) there is an mSCR that was not sufficient to produce anectopic (unstimulated) ventricular beat or arrhythmia (more than threeectopic beats). In contrast, in a heart treated with NOS1 inhibition, multiplemSCRs (see arrows) developed that were associated with multiple ectopicbeats (i.e., ventricular arrhythmia). (B) NOS1 inhibition only increased ven-tricular arrhythmias over a broad range of pacing rates under conditionsof elevated [Ca2+]i, (n = 5) compared with NOS1 inhibition with normal[Ca2+]i, (n = 5) and control hearts with normal [Ca2+]i, (n = 4), and elevated[Ca2+]i, (n = 4).

Fig. 2. Increased multicellular spontaneous Ca2+ release (mSCR) as cellular mechanism of ventricular arrhythmias with NOS1 inhibition (SMTC). (A) Exampleof a triggered arrhythmia following NOS1 inhibition. Shown are optical Ca2+ recordings from sites of earliest (red tracing) and latest (black tracing) activation.Notice that the slopes (dashed line) of the mSCRs preceding beats V1, V2, and V4 are greater in the earliest activated site compared with the latest activatedsite. Activation and mSCR slope maps confirm that the region of earliest activation (red marker) corresponds exactly to the region where mSCR slope is thegreatest for each triggered beat. (B) Example of mSCR measured in the whole heart under elevated [Ca2+]i conditions with and without NOS1 inhibition.Summary data show that following inhibition of nNOS under both normal (control, n = 4 and NOS1 inhibition, n = 5) and elevated (control, n = 4 and NOS1inhibition, n = 5) [Ca2+]i, mSCR amplitude increased and mSCR onset was faster (i.e., increased slope; decreased time to peak) compared with control hearts.*P < 0.05 compared with control.

Cutler et al. PNAS | October 30, 2012 | vol. 109 | no. 44 | 18187

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0

NOS1 Inhibition Increases Spontaneous Sarcoplasmic Reticulum Ca2+

Release and Peak L-type Ca2+. Fig. 2B shows a representative ex-ample of mSCR activity measured in the intact heart following aperiod of constant pacing, under elevated [Ca2+]i conditions withand without NOS1 inhibition. As shown, NOS1 inhibition increasedmSCR amplitude and slope and decreased the time to mSCRpeak compared with control (i.e., without NOS1 inhibition).Summary data (Fig. 2B) show that following NOS1 inhibitionmSCR amplitude increased (P < 0.05) and mSCR onset wasfaster (i.e., increased slope and decreased time to peak, P < 0.05)compared with control hearts (n = 5) under both normal (n = 4)and elevated (n = 5) [Ca2+]i conditions. The combination ofNOS1 inhibition and elevated [Ca2+]i were additive in their ef-fect on mSCR activity compared with either in isolation. Theaddition of ryanodine in a subset of control experiments eradi-cated the development of mSCRs, confirming that mSCRs areCa2+ release events from the SR, as shown previously (9).Moreover, these data show that NOS1 inhibition with or withoutelevated [Ca2+]i increase mSCR activity. However, the combi-nation of NOS1 inhibition and elevated [Ca2+]i appears to benecessary for the development of ventricular arrhythmias in theintact heart (Fig. 1B).Consistent with previous reports in isolated myocytes from the

mouse (5), we found that inhibition of NOS1 in isolated guineapig cardiomyocytes resulted in frequent spontaneous Ca2+ tran-sients and aftercontractions (100% of myocytes). In contrast,spontaneous Ca2+ transients and aftercontractions were seensignificantly less in control myocytes (22% of myocytes, P < 0.01).Moreover, inhibition of NOS1 increased peak L-type Ca2+ cur-rent (ICaL) in isolated myocytes compared with control (−9.13pA/pF vs. −5.92 pA/pF, respectively; P < 0.05). In total, thesedata are consistent with previous findings in the mouse, dem-onstrating that loss of NOS1 produces spontaneous RyR Ca2+

release and increased ICaL in isolated cardiomyocytes undernormal [Ca2+]i conditions (7).

Altered RyR2 Nitroso–Redox Balance Increases ArrhythmogenesisDuring NOS1 Inhibition. Inhibition of NOS1 can increase reactiveoxygen species (ROS), which may significantly disrupt nitroso–redox balance of key SR Ca2+ cycling proteins (e.g., RyR2).Whole heart ROS was assayed as a marker of oxidative stress andshown to increase during NOS1 inhibition compared with con-trol (Fig. 3A, P < 0.05). Importantly, treatment with the xan-thine oxidoreductase (XOR) inhibitor allopurinol reversed theROS increase during NOS1 inhibition (Fig. 3A, P < 0.05) sug-gesting that NOS1 inhibition increases XOR-derived superoxideand hydrogen peroxide. At the RyR2 level, NOS1 inhibitionresulted in significant nitroso–redox imbalance, as evidenced bydecreased RyR2-SNO [SNO-resin-assisted capture (RAC)method] as well as total RyR2 thiol content. Collectively thesedata indicate that increased RyR2-SOx (i.e., decreased RyR2thiol content) is secondary to increased XOR-derived ROS andloss of NOS1-derived NO (Fig. 3B, P < 0.05).The mechanistic role of decreased RyR2-SNO and increased

RyR2-SOx in arrhythmias was further assessed in both isolatedcardiomyocytes and the intact heart. Treatment with either al-lopurinol (XOR inhibition) or GSNO during NOS1 inhibitionsuppressed spontaneous aftercontractions in isolated myocytescompared with NOS1 inhibition alone (Fig. 4A and Fig. S1, P <0.05). Consistent with this finding, treatment with allopurinolduring NOS1 inhibition in the intact heart decreased both mSCRactivity and the incidence of ventricular arrhythmias comparedwith NOS1 inhibition alone (Fig. 4B, P < 0.01). Notably, treat-ment with either allopurinol (Fig. 3B) or GSNO (Fig. 5A) duringNOS1 inhibition increased the number of RyR2-free thiols, anindication of diminished RyR2-SOx. Also, allopurinol treatmentin the absence of NOS1 inhibition increased the number ofRyR2-free thiols compared with control (35.8 ± 1.7 vs. 31.5 ± 1.4,respectively; P < 0.01). Moreover, RyR2-SNO was increased byGSNO (Fig. 5A) treatment during NOS1 inhibition. In contrast,RyR2-SNO was not significantly increased by allopurinol treatment

during NOS1 inhibition compared with NOS1 inhibition alone(Fig. 3B). Taken together, these data indicate that oxidation andnitrosylation are antipathetic, but the effects of NO to preventoxidation cannot be explained solely by direct competition forRyR thiols. One interpretation of these data are that the effectsof NO are mediated partly through XOR inhibition (10). Thus,rather than nitrosylation and oxidation directly antagonizingeach other (e.g., competing for the same sites), the decrease innitrosylation and concomitant increase in oxidation is mediatedby XOR, which increases ROS production when local NO isdecreased (via NOS1 inhibition).Notably, treatment with NOS1 inhibitor + allopurinol (which

reduced mSCR activity and arrhythmogenesis) did not reduceICaL elevation caused by NOS1 inhibition. This suggests that theincreased ICaL is not a major mechanism for increasing arrhythmiasduring NOS1 inhibition. Previously, it has been shown that NOS1deficiency does not alter PKA-mediated RyR phosphorylation(5). However, in this study, we found that CaMKII-mediated

Fig. 3. Combination of decreased S-nitrosylation and increased oxidation ofRyR2 as molecular mechanisms of ventricular arrhythmias with NOS1 in-hibition (SMTC). (A) Whole heart optical imaging of reactive oxygen species(ROS) showed an increase in ROS during NOS1 inhibition (red line; n = 4)compared with control (black line; n = 4). Moreover, treatment with thexanthine oxidoreductase inhibitor, allopurinol during NOS1 inhibition (n = 4)decreased ROS production. (B) Mean data showed that NOS1 inhibition (n = 6)decreased RyR2-SNO and the number of RyR2-free thiols (indicating in-creased RyR2 oxidation) compared with control hearts (n = 6). Addition ofallopurinol during NOS1 inhibition (n = 6) prevented RyR2 oxidation (RyR2-SOx) but did not significantly increase RyR2-SNO. *P < 0.05 compared withcontrol. **P = 0.09 compared with control.

18188 | www.pnas.org/cgi/doi/10.1073/pnas.1210565109 Cutler et al.

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0

RyR2 phosphorylation was decreased by NOS1 inhibition com-pared with control (P < 0.05, Fig. 5B).Collectively these data support the idea that both decreased

RyR2-SNO and (resultant) increased RyR2-SOx contribute tocreate nitroso–redox imbalance, which is sufficient to increasemSCR activity and ventricular arrhythmias during NOS1 inhibition.

Severe Ventricular Arrhythmias Following Combined NOS1 Inhibitionand Oxidative Stress. To further illustrate the importance ofnitroso–redox imbalance in cardiac arrhythmogenesis and par-ticularly, the synergistic effect of SNO-deficiency and increasedoxidation, we investigated the effects of exogenous H2O2 in theintact heart with and without NOS1 inhibition. Fig. 6A showsrepresentative examples of mSCR activity following NOS1 in-hibition, H2O2 and combined NOS1 inhibition +H2O2. Both NOS1inhibition and H2O2 increased mSCR amplitude, and decreasedmSCR time to peak compared with control. Notably, the com-bination of NOS1 inhibition + H2O2 produced an additive in-crease in mSCR amplitude and decrease in mSCR time to peakcompared with NOS1 inhibition and H2O2 alone. Importantly,this increased mSCR activity resulted in a more severe, sus-tained, arrhythmia phenotype. Specifically, NOS1 inhibition +H2O2 produced ∼150% increase in the frequency of sustained(>5 s) ventricular arrhythmias compared with NOS1 inhibitionor H2O2 alone (Fig. 6B, P < 0.01).

RyR2 Nitroso–Redox Imbalance Is Not Responsible for Increase inVentricular Arrhythmias Following NOS1 Inhibition + Oxidative Stress.We postulated that combined NOS1 inhibition + H2O2 wouldproduce a greater RyR2 nitroso–redox imbalance (i.e., less freethiols) than either NOS1 inhibition or H2O2 alone (Fig. 7). BothNOS1 inhibition and H2O2 alone decreased RyR2-SNO and in-creased RyR2-SOx. However, there was no difference in RyR2-SNO and RyR2-SOx following combined NOS1 inhibition +H2O2compared with NOS1 inhibition or H2O2 alone (Fig. 7). Thus, theadditive increase in ventricular arrhythmias following combinedNOS1 inhibition + H2O2 may not be solely due to further changesin RyR2 nitroso–redox status. For example, both NOS1 inhibitionand H2O2 increased ICaL, and their combination resulted in a nearadditive increase in ICaL (P = 0.09) compared with either alone.These data suggest that in the presence of increased RyR2 Po(secondary to altered RyR2 nitroso–redox), a further increasein ICaL caused by exogenous H2O2 creates a potent substratefor arrhythmias.

DiscussionOur studies establish that diminished NOS1 activity may causearrhythmia in the intact heart. In particular, we demonstrate thatNOS1 inhibition in the intact heart increases susceptibility to Ca2+-mediated triggered arrhythmias only in presence of elevatedmyocardial [Ca2+]i. These arrhythmias arise from increasedmSCR activity (i.e., multicellular spontaneous SR Ca2+ releaseevents) due to a combination of decreased RyR2 S-nitrosylationand increased RyR2 S-oxidation. Restoration of nitroso–redoxbalance by inhibition of xanthine oxidoreductase (allopurinol) orincreasing NO bioavailability (GSNO) reversed the arrhythmo-genic phenotype. We also provide evidence for NOS1 regulationof RyR2 phosphorylation at the CaMKII site (S2814), suggestingmolecular cross-talk between the nitrosylation, oxidation, andphosphorylation states of RyR2. Finally, oxidative stress (as inischemic heart disease) may predispose to a strikingly severearrhythmic phenotype in the absence of NOS1-derived NO.

Effect of NOS1 Inhibition on Ca2+-Mediated Arrhythmias. In the presentstudy we demonstrate that NOS1 inhibition alone increases trig-gered arrhythmias but only in the presence of elevated [Ca2+]i. Thecellular mechanisms underlying these ventricular arrhythmias areprimarily from spontaneous calcium release (i.e., mSCR) thatappears to be caused by a primary alteration in RyR2 function.Also, we found that the site of earliest activation of an ectopic beatoccurred in the region with the steepest slope of mSCR activity,further suggesting a Ca2+-mediated triggered arrhythmia. This isconsistent with previous research from our laboratory, demon-strating that the rate of mSCR rise (i.e., slope) is an importantdeterminant of Ca2+-mediated triggered activity in heart prepara-tions (8). Moreover, we found that increased ICaL is not a requiredmechanism for spontaneous calcium release during NOS1 in-hibition, further supporting a primary role of altered RyR2 function.

Fig. 4. Allopurinol reverses the effects of NOS1 inhibition (SMTC) onarrhythmogenic aftercontractions in isolated myocytes and ventriculararrhythmias in the intact heart. (A) Tracings of myocyte length showing anincrease in aftercontractions with NOS1 inhibition (n = 31) compared withcontrol (n = 27). Treatment with allopurinol (n = 14) during NOS1 inhibitionsuppressed aftercontractions. (B) Tracings of allopurinol treatment duringNOS1 inhibition (n = 4) reversed the effects of NOS1 inhibition (n = 5) ontriggered arrhythmias in the intact heart to near control levels (n = 5). *P <0.05 compared with control.

Control SMTC SMTC +

GSNO

Control SMTC SMTC +

GSNO

A B

Fig. 5. Treatment with GSNO and CaMKII mediated RyR2 phosphorylation.(A) Treatment with GSNO during NOS1 inhibition decreased RyR2-SOx andincreased RyR2-SNO compared with NOS1 inhibition alone. (B) Representa-tive Western blots of RyR2-PS2814/RyR2Total during NOS1 inhibition comparedwith control. As shown, NOS1 inhibition decreased RyR2-PS2814 comparedwith control. *P < 0.05 compared with control.

Cutler et al. PNAS | October 30, 2012 | vol. 109 | no. 44 | 18189

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0

These findings extend the observations of Gonzalez et al. (5).in isolated myocytes to the realm of system physiology and dis-ease. Gonzalez showed that deletion of NOS1 increased SR Ca2+

leak and diastolic Ca2+ waves. Our data establish in the intactheart that NOS1-mediated alterations in RyR2 function are notsufficient to produce arrhythmias. For arrhythmias to developwith NOS1 inhibition, elevated myocardial [Ca2+] (a character-istic of disease) is necessary. These data highlight that the merepresence of spontaneous SR Ca2+ release events in a cardiomyocytedoes not necessarily translate into a triggered beat or arrhythmiain coupled myocardium and provide unique insight into the basisof arrhythmias in disease.

Molecular Mechanisms Driving Ca2+-Mediated Arrhythmogenesis withNOS1 Inhibition.The molecular mechanisms regulating spontaneouscalcium release in the intact heart are complex. Posttranslationalmodifications of RyR2 by S-nitrosylation and/or S-oxidation havebeen shown to affect SR Ca2+ leak and sparks in cardiomyocytes(5, 11). In the current study we demonstrate that the combination ofreduced RyR2-SNO and increased RyR2-SOx increase spontane-ous calcium release and ventricular arrhythmias in the intact heart.Importantly, these arrhythmias could be reversed with the xanthineoxidoreductase inhibitor allopurinol or the NO donor GSNO.

These data are consistent with a model in which RyR2-SNO reg-ulates RyR2-SOx, and where SR nitroso–redox imbalance underliesarrhythmogenesis. However, it is possible that other mechanismsare involved in NO-deficiency–mediated arrhythmogenesis.Notably, we demonstrated that NOS1 inhibition also increased

ICaL and decreased RyR2 phosphorylation at the CaMKII bind-ing site (S2814). ICaL is not primarily responsible for increasingarrhythmia because allopurinol suppressed arrhythmia withoutdecreasing ICaL, and despite the importance of CaMKII signalingin abnormal Ca2+ cycling (12), the significance of decreasedphosphorylation is unclear. Recently, Respress et al. (13) showeda reduction in RyR2-P2814 in mice with ischemic cardiomyopathy,a condition known to have increased RyR2 Ca2+ leak. However,S2814 mutant mice did not show increased RyR2 Ca2+ leak. Ourfindings provide unique evidence that NOS1 is involved in theregulation of RyR2-P2814 and suggest cross-talk between thenitrosylation and phosphorylation states of proteins (i.e., RyR2).This suggests an additional level of signaling complexity in post-translational regulation of RyR2.The mechanisms responsible for reduced RyR2-P2814 during

NOS1 inhibition are likely complex. It could be postulated thateither alone or in combination decreased RyR2-SNO or increasedRyR2-SOx produces a conformational change in RyR2, such thatthe S2814 phosphorylation site is less available for phosphory-lation. Alternatively, it has been shown that S-nitrosylation ofphosphatases can inhibit their activity (14). Thus, it is possiblethat NOS1 inhibition decreases S-nitrosylation of a key proteinphosphatase (e.g., PP2A); in turn reducing RyR2 phosphorylationat the CaMKII site (S2814).Importantly, our findings may explain differences between

Gonzalez et al. (5) and Wang et al. (6) regarding the effect NOS1deficiency on RyR Ca2+ leak and Ca2+ sparks/waves. Gonzalezet al. (5) reported increased Ca2+ wave frequency in cardiomyocytesfrom NOS1-KO mice whereas Wang et al. (6) reported a de-crease in Ca2+ spark frequency. In studies by us and Gonzalezet al. (5) rapid pacing was performed to induce Ca2+ waves andarrhythmias, whereas very slow pacing was used by Wang et al.(6). Pacing can alter the redox state of the cardiomyocyte orincrease SR [Ca2+], thereby altering the probability of Ca2+

sparks/waves.

Fig. 6. Combined NOS1 inhibition (SMTC) and exogenous H2O2 produce anadditive increase in mSCR and sustained (>5 s) triggered ventriculararrhythmias in the intact heart. (A) Example of spontaneous (nonelectricallydriven) calcium release occurring in multiple, neighboring cells under allexperimental conditions. Both H2O2 (n = 4) and NOS1 inhibition (n = 7) in-creased mSCR amplitude, and time to mSCR peak decreased compared withcontrol. Moreover, combined NOS1 inhibition and H2O2 (n = 6) produced anadditive increase in mSCR activity. (B) H2O2, NOS1 inhibition, and combinedNOS1 inhibition and H2O2 all increased the frequency of nonsustained (≤5 s)triggered arrhythmias compared with control. In contrast, only combinedNOS1 inhibition and H2O2 increased the frequency of sustained (>5 s) trig-gered arrhythmias. *P < 0.05 compared with control. **P < 0.05 comparedwith control, NOS1 inhibition and H2O2.

Fig. 7. Altered RyR2 nitroso–redox balance not responsible for additiveincrease in ventricular arrhythmias following combined NOS1 inhibition(SMTC) + exogenous H2O2. (Upper) SNO-RAC showing decreased RyR2-SNOwith NOS1 inhibition, H2O2, and NOS1 inhibition + H2O2 compared withcontrol. (Lower) Mean data showed that both NOS1 inhibition (n = 6) andH2O2 (n = 6) alone decreased RyR2-SNO and decreased RyR2-free thiols (i.e.,increased RyR2-SOx). Interestingly, there was no difference in RyR2-SNO andRyR2-SOx following combined NOS1 inhibition + H2O2 (n = 6) when com-pared with NOS1 inhibition or oxidative stress alone. *P < 0.05 comparedwith control. **P = 0.06 compared with control. P̂ = 0.08.

18190 | www.pnas.org/cgi/doi/10.1073/pnas.1210565109 Cutler et al.

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0

Combination of Decreased NOS1 Activity and Oxidative Stress MarkedlyIncrease Arrhythmogenesis. Oxidative stress associated with diseasehas been shown to increase the frequency of Ca2+ sparks, after-depolarizations, and ventricular arrhythmias. Saraiva et al. (15)reported an increased mortality in NOS1 knockout (KO) micefollowing MI and Burger et al. (7) reported both an increasedmortality and incidence of ventricular arrhythmias post-MI (leftanterior descending coronary artery ligation) in NOS1-KO mice.Importantly, many of these arrhythmias occurred within 1 h ofspontaneous death. Burger et al. (7) postulated that this increasein arrhythmogenesis was related to increased ICaL as both [Ca2+]Ielevation and ventricular arrhythmias were decreased with ve-rapamil. However, they measured ICaL in isolated cardiomyocytesfrom hearts without MI and did not assess RyR2-SOx or otherposttranslational modifications.To simulate disease we imposed oxidative stress with H2O2 and

found that NO deficiency produced a striking increase in mSCRsand ventricular arrhythmias. Interestingly, the present study showslittle qualitative difference in RyR2 oxidation state (net RyR2nitroso–redox state) with the combination of NOS1 inhibition andH2O2 versus either alone. This implies that nitroso–redox imbalancewithin RyR2 is not solely responsible for the severe arrhythmias wereport. Our methods do not exclude the possibility, however, thatthe populations of thiols oxidized under various conditions may bedifferent. Additionally, NOS1 inhibition and H2O2 showed syner-gistic effects on ICaL, and it is possible, that this synergistic increasein ICaL may in part explain the increased probability and severity ofspontaneous calcium release and arrhythmias.

Clinical Significance. Our findings have important clinical impli-cations as they highlight the role NO derived from NOS1 in thegenesis of triggered arrhythmias and identify a potential therapeuticapproach (i.e., xanthine oxidoreductase inhibition in combina-tion with NO-based therapy). There is evidence to suggest thatobesity, diabetes, and heart failure are all associated with de-creased NOS1 activity (16–20). Junttila et al. (21) recently reporteda markedly increased incidence of sudden arrhythmic death indiabetic patients after MI compared with nondiabetic controls.Diabetic patients with left ventricle ejection fractions (LVEF)>35% had a sudden death rate equal to nondiabetic patientswith LVEF <35%. It could be postulated that this increasedarrhythmic risk in diabetic patients is secondary to the combinedeffects of oxidative stress (i.e., ischemic cardiomyopathy) andimpaired NOS1 activity (i.e., diabetes). Similarly, there is pre-liminary evidence that xanthine oxidoreductase inhibition inheart failure patients with elevated uric acid improved clinical

outcomes (composite of heart failure morbidity, mortality, andquality of life) (22). Our findings suggest that therapies targetingthe restoration of nitroso–redox balance in the heart could preventsudden arrhythmic death.

Materials and MethodsWhole Heart Studies. Experiments were performed in accordance with theInstitutional Animal Care and Use Committee of Case Western ReserveUniversity. Hearts from guinea pigs were Langendorff perfused with oxy-genated (100% O2) Tyrode’s solution (pH 7.45, 34 °C). Endocardial cryoa-blations were performed to eliminate the Purkinje fibers and deeper layersof myocardium. Indo-1 AM (10 μM; Invitrogen) was used to measure in-tracellular Ca2+ transients from the anterior surface of the heart in the fol-lowing groups: (i) control, normal [Ca2+]I; (ii) NOS1 inhibition (S-methyl-L-thiocitrulline), normal [Ca2+]I; (iii) control, elevated [Ca2+]i (5.5 mM); (iv)NOS1 inhibition, elevated [Ca2+]I; (v) NOS1 inhibition + allopurinol (elevated[Ca2+]i); (vi) H2O2 (elevated [Ca2+]i) and NOS1 inhibition + H2O2 (elevated[Ca2+]i). Constant pacing for 1 min followed by a halt in pacing was used toelicit mSCR activity and arrhythmias as previously described (9). In separatehearts, the concentration of ROS in the intact heart was measured by per-fusing for 15 min with the ROS-sensitive dye dihydrorhodamine 6G (Invi-trogen). For further information, see SI Materials and Methods.

Isolated Myocytes. Myocytes were isolated from guinea pig hearts and sar-comere shortening, intracellular Ca2+ transients, and the L-type Ca2+ currentwere measured. For further information, see SI Materials and Methods.

Biochemistry. Tomeasure RyR2 S-nitrosylation, SNO-RAC was performed in SRvesicles isolated from guinea pig whole heart homogenate as describedpreviously (23). Quantification of RyR2-free thiols was also assessed in SRvesicles that were incubated with the fluorescent indicator monobromobimane(MBB) (24). Finally, RyR2 phosphorylation (RyR2-PS2814) was measured in tissueobtained from the left ventricular free wall in control and NOS1-inhibitedhearts. For further information, see SI Materials and Methods.

Statistical Analysis. Statistical differences were assessed with ANOVA, Studentt test, Kruskal–Wallis test, and χ2 test as appropriate. When unequal variancewas detected, the Wilcoxon rank sum test was used. Results were expressedas mean ± SEM.

ACKNOWLEDGMENTS. This study was supported by National Institutes ofHealth (NIH) Grant HL-054807 (to D.S.R.), NIH Grants HL-084142 and HL-100105 (to K.R.L.), NIH Grants HL-075443, HL-095463, AI-080633, and DARPAN66001-10-C-2015 (to J.S.S.), and fellowship awards from the Heart RhythmSociety, NIH National Research Service Award, and the Department of MedicinePhysician Scientist Pathway at the MetroHealth Campus of Case WesternReserve University (to M.J.C.).

1. Tamargo J, Caballero R, Gómez R, Delpón E (2010) Cardiac electrophysiological effectsof nitric oxide. Cardiovasc Res 87(4):593–600.

2. Lima B, Forrester MT, Hess DT, Stamler JS (2010) S-nitrosylation in cardiovascularsignaling. Circ Res 106(4):633–646.

3. Sun J, et al. (2008) Regulation of the cardiac muscle ryanodine receptor by O(2)tension and S-nitrosoglutathione. Biochemistry 47(52):13985–13990.

4. Stamler JS, Sun QA, Hess DT (2008) A SNO storm in skeletal muscle. Cell 133(1):33–35.5. Gonzalez DR, Beigi F, Treuer AV, Hare JM (2007) Deficient ryanodine receptor S-ni-

trosylation increases sarcoplasmic reticulum calcium leak and arrhythmogenesis incardiomyocytes. Proc Natl Acad Sci USA 104(51):20612–20617.

6. Wang H, et al. (2010) Regulation of myocyte contraction via neuronal nitric oxidesynthase: Role of ryanodine receptor S-nitrosylation. J Physiol 588(Pt 15):2905–2917.

7. BurgerDE, et al. (2009)Neuronalnitric oxide synthaseprotects againstmyocardial infarction-induced ventricular arrhythmia and mortality in mice. Circulation 120(14):1345–1354.

8. Hoeker GS, Katra RP, Wilson LD, Plummer BN, Laurita KR (2009) Spontaneous calciumrelease in tissue from the failing canine heart. Am J Physiol Heart Circ Physiol 297(4):H1235–H1242.

9. Plummer BN, Cutler MJ, Wan X, Laurita KR (2011) Spontaneous calcium oscillationsduring diastole in the whole heart: The influence of ryanodine reception function andgap junction coupling. Am J Physiol Heart Circ Physiol 300(5):H1822–H1828.

10. Khan SA, et al. (2004) Neuronal nitric oxide synthase negatively regulates xanthineoxidoreductase inhibition of cardiac excitation-contraction coupling. Proc Natl AcadSci USA 101(45):15944–15948.

11. Gonzalez DR, Treuer AV, Castellanos J, Dulce RA, Hare JM (2010) Impaired S-nitro-sylation of the ryanodine receptor caused by xanthine oxidase activity contributes tocalcium leak in heart failure. J Biol Chem 285(37):28938–28945.

12. Bers DM (2012) Ryanodine receptor S2808 phosphorylation in heart failure: Smokinggun or red herring. Circ Res 110(6):796–799.

13. Respress JL, et al. (2012) Role of RyR2 phosphorylation at S2814 during heart failureprogression. Circ Res 110(11):1474–1483.

14. Numajiri N, et al. (2011) On-off system for PI3-kinase-Akt signaling through S-nitro-sylation of phosphatase with sequence homology to tensin (PTEN). Proc Natl Acad SciUSA 108(25):10349–10354.

15. Saraiva RM, et al. (2005) Deficiency of neuronal nitric oxide synthase increases mor-tality and cardiac remodeling after myocardial infarction: Role of nitroso-redoxequilibrium. Circulation 112(22):3415–3422.

16. Umar S, van der Laarse A (2010) Nitric oxide and nitric oxide synthase isoforms in thenormal, hypertrophic, and failing heart. Mol Cell Biochem 333(1–2):191–201.

17. Koo JR, Vaziri ND (2003) Effects of diabetes, insulin and antioxidants on NO synthaseabundance and NO interaction with reactive oxygen species. Kidney Int 63(1):195–201.

18. Yu WJ, et al. (2000) Insulin restores neuronal nitric oxide synthase expression instreptozotocin-induced diabetic rats. Life Sci 68(6):625–634.

19. Silberman GA, et al. (2010) Uncoupled cardiac nitric oxide synthase mediates diastolicdysfunction. Circulation 121(4):519–528.

20. Saraiva RM, et al. (2007) Reduced neuronal nitric oxide synthase expression contrib-utes to cardiac oxidative stress and nitroso-redox imbalance in ob/ob mice. NitricOxide 16(3):331–338.

21. Junttila MJ, et al. (2010) Sudden cardiac death after myocardial infarction in patientswith type 2 diabetes. Heart Rhythm 7(10):1396–1403.

22. Hare JM, et al.; OPT-CHF Investigators (2008) Impact of oxypurinol in patients withsymptomatic heart failure. Results of the OPT-CHF study. J Am Coll Cardiol 51(24):2301–2309.

23. Forrester MT, et al. (2009) Proteomic analysis of S-nitrosylation and denitrosylation byresin-assisted capture. Nat Biotechnol 27(6):557–559.

24. Kosower NS, Kosower EM (1987) Thiol labeling with bromobimanes. Methods Enzymol143:76–84.

Cutler et al. PNAS | October 30, 2012 | vol. 109 | no. 44 | 18191

PHYS

IOLO

GY

Dow

nloa

ded

by g

uest

on

Apr

il 21

, 202

0